|
Abstract:
In this talk I will describe the MouseFunc project where the goal is to evaluate various computational methods to predict mouse gene functions using genomic data from heterogeneus sources. Several algorithms using diverse genomic data have been applied to gene function prediction, but these have been primarily tested on the unicellular yeast S. cerevisiae. To facilitate prediction of mouse gene function, identify strengths and weaknesses of existing functional genomic datasets and compare algorithm performance, we assembled a collection of genomic data on the mouse M. Musculus. Nine bioinformatics teams used this dataset to independently train classifiers and generate predictions of function for 21,603 mouse genes. Performance between the teams was assessed using both a held-out test set of genes and a prospective evaluation of novel predictions.